OpenCL

timchist · ‎11-30-2022

Earlier versions of AMD drivers (for example 19.12.3 and many others) reported RX 550 as gfx904. Newer versions (such as 22.11.1) report this GPU as gfx903. I could not find a single driver that is capable of producing binaries for gfx903 on Windows when compiling OpenCL code in offline mode.

How can I prepare a binary for OpenCL code working with RX 550 when recent drivers are installed?

timchist · ‎02-01-2023

Fiji binary seems to be working indeed.

It also appears that the gfx804 binary also works. Could you please confirm with the OpenCL team -- which one is better to use: Fiji or gfx804?

View solution in original post

dipak · ‎02-21-2023

As the OpenCL team suggested, you should use "Fiji" offline binary for all gfx803 GPUs.

Thanks.

View solution in original post

dipak · ‎12-02-2022

Hi @timchist ,

Thank you for the query. I have forwarded it to the OpenCL team. I will let you know once I get their feedback on this.

Thanks.

dipak · ‎12-09-2022

As I have been informed, the OpenCL team ran some related tests with the latest internal driver build, but they didn't observe any issue. Could you please share your clinfo output?

Thanks.

timchist · ‎12-12-2022

Adrenalin 19.12.3:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3004.7)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFE02E1BFD0
  Name:						 gfx804
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3004.7
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3004.7)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv

Adrenalin 22.11.1:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3444.0)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFB75648000
  Name:						 gfx803
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3444.0 (PAL,HSAIL)
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3444.0)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_copy_buffer_p2p cl_amd_planar_yuv

When I try to compile OpenCL code for gfx803 offline (using CL_CONTEXT_OFFLINE_DEVICES_AMD), none of the drivers I have tried is able to produce binary. Old drivers can do gfx804, but not gfx803.

timchist · ‎12-12-2022

Adrenalin 19.12.3:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3004.7)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFE02E1BFD0
  Name:						 gfx804
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3004.7
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3004.7)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv

Adrenalin 22.11.1:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3444.0)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFB75648000
  Name:						 gfx803
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3444.0 (PAL,HSAIL)
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3444.0)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_copy_buffer_p2p cl_amd_planar_yuv

Older Windows drivers can produce binaries for gfx804 (when using CL_CONTEXT_OFFLINE_DEVICES_AMD flag), but no drivers can generate a binary for gfx803.

timchist · ‎12-12-2022

Adrenalin 19.12.3:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3004.7)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFE02E1BFD0
  Name:						 gfx804
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3004.7
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3004.7)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv

Adrenalin 22.11.1:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3444.0)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFB75648000
  Name:						 gfx803
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3444.0 (PAL,HSAIL)
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3444.0)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_copy_buffer_p2p cl_amd_planar_yuv

Older Windows drivers can produce binaries for gfx804 (when using CL_CONTEXT_OFFLINE_DEVICES_AMD flag), but no drivers can generate a binary for gfx803.

timchist · ‎12-12-2022

Adrenalin 19.12.3:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3004.7)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFE02E1BFD0
  Name:						 gfx804
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3004.7
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3004.7)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv

Adrenalin 22.11.1:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3444.0)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFB75648000
  Name:						 gfx803
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3444.0 (PAL,HSAIL)
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3444.0)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_copy_buffer_p2p cl_amd_planar_yuv

Older Windows drivers can produce binaries for gfx804 (when using CL_CONTEXT_OFFLINE_DEVICES_AMD flag), but no drivers can generate a binary for gfx803.

timchist · ‎12-12-2022

The message I post with output from clinfo from different drivers disappears for unclear reason. Please advise.

timchist · ‎12-12-2022

Adrenalin 19.12.3:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3004.7)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFE02E1BFD0
  Name:						 gfx804
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3004.7
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3004.7)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv

dipak · ‎12-12-2022

The post was wrongly marked as spam by the community auto-spam filer. I have restored it.

Thanks.

timchist · ‎12-12-2022

Adrenalin 19.12.3:

Number of platforms: 1
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 2.1 AMD-APP (3004.7)
Platform Name: AMD Accelerated Parallel Processing
Platform Vendor: Advanced Micro Devices, Inc.
Platform Extensions: cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices

Platform Name: AMD Accelerated Parallel Processing
Number of devices: 1
Device Type: CL_DEVICE_TYPE_GPU
Vendor ID: 1002h
Board name: Radeon RX550/550 Series
Device Topology: PCI[ B#1, D#0, F#0 ]
Max compute units: 8
Max work items dimensions: 3
Max work items[0]: 1024
Max work items[1]: 1024
Max work items[2]: 1024
Max work group size: 256
Preferred vector width char: 4
Preferred vector width short: 2
Preferred vector width int: 1
Preferred vector width long: 1
Preferred vector width float: 1
Preferred vector width double: 1
Native vector width char: 4
Native vector width short: 2
Native vector width int: 1
Native vector width long: 1
Native vector width float: 1
Native vector width double: 1
Max clock frequency: 1071Mhz
Address bits: 64
Max memory allocation: 1597190963
Image support: Yes
Max number of images read arguments: 128
Max number of images write arguments: 64
Max image 2D width: 16384
Max image 2D height: 16384
Max image 3D width: 2048
Max image 3D height: 2048
Max image 3D depth: 2048
Max samplers within kernel: 16
Max size of kernel argument: 1024
Alignment (bits) of base address: 2048
Minimum alignment (bytes) for any datatype: 128
Single precision floating point capability
Denorms: No
Quiet NaNs: Yes
Round to nearest even: Yes
Round to zero: Yes
Round to +ve and infinity: Yes
IEEE754-2008 fused multiply-add: Yes
Cache type: Read/Write
Cache line size: 64
Cache size: 16384
Global memory size: 2147483648
Constant buffer size: 1597190963
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 32768
Max pipe arguments: 16
Max pipe active reservations: 16
Max pipe packet size: 1597190963
Max global variable size: 1437471744
Max global variable preferred total size: 2147483648
Max read/write image args: 64
Max on device events: 1024
Queue on device max size: 8388608
Max on device queues: 1
Queue on device preferred size: 262144
SVM capabilities:
Coarse grain buffer: Yes
Fine grain buffer: Yes
Fine grain system: No
Atomics: No
Preferred platform atomic alignment: 0
Preferred global atomic alignment: 0
Preferred local atomic alignment: 0
Kernel Preferred work group size multiple: 64
Error correction support: 0
Unified memory for Host and Device: 0
Profiling timer resolution: 1
Device endianess: Little
Available: Yes
Compiler available: Yes
Execution capabilities:
Execute OpenCL kernels: Yes
Execute native function: No
Queue on Host properties:
Out-of-Order: No
Profiling : Yes
Queue on Device properties:
Out-of-Order: Yes
Profiling : Yes
Platform ID: 00007FFE02E1BFD0
Name: gfx804
Vendor: Advanced Micro Devices, Inc.
Device OpenCL C version: OpenCL C 2.0
Driver version: 3004.7
Profile: FULL_PROFILE
Version: OpenCL 2.0 AMD-APP (3004.7)
Extensions: cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv

timchist · ‎12-12-2022

Adrenalin 19.12.3:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3004.7)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFE02E1BFD0
  Name:						 gfx804
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3004.7
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3004.7)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv

timchist · ‎12-12-2022

Adrenalin 22.11.1:

Number of platforms:				 1
  Platform Profile:				 FULL_PROFILE
  Platform Version:				 OpenCL 2.1 AMD-APP (3444.0)
  Platform Name:				 AMD Accelerated Parallel Processing
  Platform Vendor:				 Advanced Micro Devices, Inc.
  Platform Extensions:				 cl_khr_icd cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_event_callback cl_amd_offline_devices 


  Platform Name:				 AMD Accelerated Parallel Processing
Number of devices:				 1
  Device Type:					 CL_DEVICE_TYPE_GPU
  Vendor ID:					 1002h
  Board name:					 Radeon RX550/550 Series
  Device Topology:				 PCI[ B#1, D#0, F#0 ]
  Max compute units:				 8
  Max work items dimensions:			 3
    Max work items[0]:				 1024
    Max work items[1]:				 1024
    Max work items[2]:				 1024
  Max work group size:				 256
  Preferred vector width char:			 4
  Preferred vector width short:			 2
  Preferred vector width int:			 1
  Preferred vector width long:			 1
  Preferred vector width float:			 1
  Preferred vector width double:		 1
  Native vector width char:			 4
  Native vector width short:			 2
  Native vector width int:			 1
  Native vector width long:			 1
  Native vector width float:			 1
  Native vector width double:			 1
  Max clock frequency:				 1071Mhz
  Address bits:					 64
  Max memory allocation:			 1597190963
  Image support:				 Yes
  Max number of images read arguments:		 128
  Max number of images write arguments:		 64
  Max image 2D width:				 16384
  Max image 2D height:				 16384
  Max image 3D width:				 2048
  Max image 3D height:				 2048
  Max image 3D depth:				 2048
  Max samplers within kernel:			 16
  Max size of kernel argument:			 1024
  Alignment (bits) of base address:		 2048
  Minimum alignment (bytes) for any datatype:	 128
  Single precision floating point capability
    Denorms:					 No
    Quiet NaNs:					 Yes
    Round to nearest even:			 Yes
    Round to zero:				 Yes
    Round to +ve and infinity:			 Yes
    IEEE754-2008 fused multiply-add:		 Yes
  Cache type:					 Read/Write
  Cache line size:				 64
  Cache size:					 16384
  Global memory size:				 2147483648
  Constant buffer size:				 1597190963
  Max number of constant args:			 8
  Local memory type:				 Scratchpad
  Local memory size:				 32768
  Max pipe arguments:				 16
  Max pipe active reservations:			 16
  Max pipe packet size:				 1597190963
  Max global variable size:			 1437471744
  Max global variable preferred total size:	 2147483648
  Max read/write image args:			 64
  Max on device events:				 1024
  Queue on device max size:			 8388608
  Max on device queues:				 1
  Queue on device preferred size:		 262144
  SVM capabilities:				 
    Coarse grain buffer:			 Yes
    Fine grain buffer:				 Yes
    Fine grain system:				 No
    Atomics:					 No
  Preferred platform atomic alignment:		 0
  Preferred global atomic alignment:		 0
  Preferred local atomic alignment:		 0
  Kernel Preferred work group size multiple:	 64
  Error correction support:			 0
  Unified memory for Host and Device:		 0
  Profiling timer resolution:			 1
  Device endianess:				 Little
  Available:					 Yes
  Compiler available:				 Yes
  Execution capabilities:				 
    Execute OpenCL kernels:			 Yes
    Execute native function:			 No
  Queue on Host properties:				 
    Out-of-Order:				 No
    Profiling :					 Yes
  Queue on Device properties:				 
    Out-of-Order:				 Yes
    Profiling :					 Yes
  Platform ID:					 00007FFB75648000
  Name:						 gfx803
  Vendor:					 Advanced Micro Devices, Inc.
  Device OpenCL C version:			 OpenCL C 2.0 
  Driver version:				 3444.0 (PAL,HSAIL)
  Profile:					 FULL_PROFILE
  Version:					 OpenCL 2.0 AMD-APP (3444.0)
  Extensions:					 cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_copy_buffer_p2p cl_amd_planar_yuv

Older Windows drivers can produce binaries for gfx804 (when using CL_CONTEXT_OFFLINE_DEVICES_AMD flag), but no drivers can generate a binary for gfx803.

dipak · ‎12-13-2022

Thanks for providing the clinfo output.

It seems like the issue has been fixed in the latest internal driver. The version of the internal driver, which was used to run the tests, is higher than this one. So, please wait till the fix is available with public driver.

Thanks.

dipak · ‎12-14-2022

>>Older Windows drivers can produce binaries for gfx804 (when using CL_CONTEXT_OFFLINE_DEVICES_AMD flag), but no drivers can generate a binary for gfx803.

Sorry, I missed the above point. In the first post, you mentioned the offline binary for gfx903, so I checked with the OpenCL team for gfx903.

timchist · ‎12-18-2022

I need to either be able to generate a binary for gfx903 (using any version of a driver), or have the newer drivers updated so that they report this GPU as gfx904, as it used to be before.

At the moment there is no way to generate a binary for RX 550 offline that could be used with a new driver.

Is there a solution?

dipak · ‎12-19-2022

From the above clinfo information, it looks like older Adrenalin drivers (e.g. 19.12.3 ) used to report RX550 as gfx804 whereas newer Adrenalin drivers (e.g. 22.11.1) reports it as gfx803.

So, could you please clarify the statement below? Is it gfx904 or gfx804?

>> I need to either be able to generate a binary for gfx903 (using any version of a driver), or have the newer drivers updated so that they report this GPU as gfx904, as it used to be before.

If it's related to gfx804, then I need to check with the OpenCL team.

Thanks.

timchist · ‎12-25-2022

Sorry for the confusion dipak.

Old drivers reported the GPU as gfx804.

New drivers report it as gfx803.

I need a way to compile offline for this GPU working with newer drivers.

dipak · ‎12-26-2022

Thanks for the clarification. I have reported it to the OpenCL team. Once I get any feedback from them, I will let you know.

Thanks.

dipak · ‎01-05-2023

A bug ticket has been created to investigate this issue.

Thanks.

dipak · ‎01-27-2023

Hi @timchist ,

Below is the suggestion from the OpenCL team:

"We suggest the developer to use Fiji offline device to generate binary for all gfx803 GPUs. That is, use offline device which clGetDeviceInfo ( CL_DEVICE_NAME ) returns device name "Fiji".

So, please try the following step and let us know if it works for you: generate offline binary for "Fiji" and then, load and run it on RX550 (or gfx803).

Thanks.

timchist · ‎02-01-2023

Thanks dipak. We will give it a try.

timchist · ‎02-01-2023

Fiji binary seems to be working indeed.

It also appears that the gfx804 binary also works. Could you please confirm with the OpenCL team -- which one is better to use: Fiji or gfx804?

dipak · ‎02-01-2023

Thanks for the confirmation.

>>It also appears that the gfx804 binary also works. Could you please confirm with the OpenCL team -- which one is better to use: Fiji or gfx804?

As you mentioned earlier, old drivers used to report RX550 as gfx804. So, are you referring to the offline binary generated with old drivers for gfx804?

Thanks.

timchist · ‎02-05-2023

> So, are you referring to the offline binary generated with old drivers for gfx804?

Yes.

dipak · ‎02-05-2023

Okay, I will check with the OpenCL team and get back to you.

Thanks.

dipak · ‎02-21-2023

As the OpenCL team suggested, you should use "Fiji" offline binary for all gfx803 GPUs.

Thanks.

OpenCL

Cannot offline compile OpenCL code for RX 550 in new drivers on Windows